[PATCH] #3807: 'svn up' of a nonexistent child in a copied dir triggers an assertion

[Daniel Shahaf <da...@elego.de>]

http://subversion.tigris.org/issues/show_bug.cgi?id=3807

I've had a look at the above, and I'm attaching a patch.  I'd like to get
review of this patch --- I have not worked enough on wc internals to feel
confident about it, and I wrote at least part of it by an "Add an if() until
the tests stop segfault" approach.

Thanks.

Daniel

Re: [PATCH] #3807: 'svn up' of a nonexistent child in a copied dir triggers an assertion

[Philip Martin <ph...@wandisco.com>]

Daniel Shahaf <da...@elego.de> writes:

> ###: what *should* the 'svn up' do?  Should it always succeed
> silently, or should

> ### 'svn cp A A2; svn mkdir ^/A/foo; svn up A2/foo' pull the new
> directory?

I don't think it should pull the new foo.  At present "svn cp" fixes the
copyfrom revision and a subsequent "svn up" doesn't change it.

Having update change the copyfrom revision does look like a useful
feature, it's a way to resolve some tree conflicts, but I think we
should try for a full implementation rather than just a few corner
cases.

-- 
Philip

Re: [PATCH] #3807: 'svn up' of a nonexistent child in a copied dir triggers an assertion

[Daniel Shahaf <da...@elego.de>]

Committed r1078497.